Re: [core] Review draft-tiloca-core-oscore-discovery-04

[Christian M. Amsüss <[e-mail]>]

Hello Jim,

> Need to go here as well.
This too.

just picking the points I feel I have something to say about:

> * Should this document be re-written to only use CoRAL-reef?

I don't think so in general for time line reasons. (There's no concrete
pressure on this as far as I know of, but CoRAL still has no forseeable
timeline).

What I could see us do is having both in examples, and possibly even
accept that some features are not as optimized or usable in link-format
but can be in CoRAL. (The link from the join resource to the
authorization server, on which there was the comment today from
proxy-Klaus, comes to mind.)

> * Section 5 - I don't care if you specify the same application group
> multiple times.   I don't want to have to figure out how to check this one
> thing and error.  Plus the end result will always be the same anyway

Considering the paragraph below that, the app-gp paragraph can be vastly
simplified anyway, and I think we already have an item about making this
less normative. (The link's content is normative; how it's looked up in
an RD is just a description of the combined steps).

> * Section 5.1 - I think the anchor in the response is supposed to be absent.

The anchor needs to be there because it's an RD lookup result, and the
RD needs to fill in the anchor because it could be relevant for
interpreting the implied "hosts" relation or any non-target link
attributes (which I usually claim should not be registered, but that's
what we arrived at RD for 6690 compatibility).

Thanks for your input
c

-- 
To use raw power is to make yourself infinitely vulnerable to greater powers.
  -- Bene Gesserit axiom